\begin{table}

\caption{Regression Models: Heterogenous effects of the Deactivation conditional on WhatsApp Usage for Politics }
\centering
\begin{threeparttable}
\begin{tabular}[t]{lcccccc}
\toprule
  & False Rumors Exposure & True News Exposure & False Rumors Accuracy & True News Accuracy & Polarization Index & Subjective Well-Being Index\\
\midrule
Treatment & -0.324* & -0.299** & -0.285* & -0.133 & -0.037 & 0.236\\
 & (0.139) & (0.102) & (0.143) & (0.135) & (0.259) & (0.470)\\
WhatsApp usage for Politics & 0.034 & 0.058+ & -0.092* & 0.021 & -0.073 & -0.225+\\
 & (0.042) & (0.034) & (0.041) & (0.039) & (0.077) & (0.121)\\
Treatment x WhatsApp usage for Politics & -0.073 & -0.011 & 0.097* & 0.029 & 0.038 & 0.085\\
 & (0.051) & (0.040) & (0.048) & (0.048) & (0.097) & (0.159)\\
\midrule
Num.Obs. & 518 & 517 & 518 & 517 & 517 & 517\\
R2 & 0.135 & 0.126 & 0.192 & 0.111 & 0.145 & 0.151\\
R2 Adj. & 0.082 & 0.072 & 0.142 & 0.057 & 0.092 & 0.099\\
RMSE & 0.99 & 0.84 & 1.00 & 0.98 & 1.97 & 3.24\\
\bottomrule
\multicolumn{7}{l}{\rule{0pt}{1em}+ p $<$ 0.1, * p $<$ 0.05, ** p $<$ 0.01, *** p $<$ 0.001}\\
\end{tabular}
\begin{tablenotes}
\item \textit{Note: } 
\item  Robust standard errors in Parentheses. All models use the Covariate-Adjusted ITT estimator. 
\end{tablenotes}
\end{threeparttable}
\end{table}
